My Dream Auto Trading Software (repost from Elite Trader)

Questions about MultiCharts and user contributed studies.
flipflopper
Posts: 261
Joined: 28 Feb 2008
Has thanked: 2 times
Been thanked: 1 time

My Dream Auto Trading Software (repost from Elite Trader)

Postby flipflopper » 04 Feb 2010

1. Has the ability to trade a portfolio of instruments.
a) Multicharts allows me to backtest and optimize a portfolio but when it comes to trading I have to add each individual instrument and apply the strategy to it. This makes it very difficult for the system as a whole to know what the other strategies are doing (whether they are in a position or not how much capital is at risk at any given time). What I consider to be the true holy grail is the ability to efficiently use capital at all times while never risking more then a certain percentage at any given time. Without the ability to create strategies on the portfolio level it is impossible to know what is going on in the big picture at any given time.
b) The strategy must always know what your current intraday and overnight buying power is so it can adjust positions to meet overnight requirements but is also able to take advantage of additional intraday margin.

2. Can more accurately calculate slippage in back testing based on the bid ask spread, the avg daily volume, and the size of your position.

3. When backtesting and optimizing a portfolio of instruments the results should allow you to drill down and see how each individual instrument performed. Also to group instruments that performed similarly together so you can easily identify which group of instruments works well with the strategy.

4. The ability to optimize based on Sharpe Ratio, Lowest Drawdown, least consecutive losers, etc. NetProfit is not the most important factor.

I would be interested to hear from any other auto traders on their opinions and thoughts. Does something like this exist? Would you have to build it? Is this stuff useful or not really?

User avatar
geizer
Posts: 375
Joined: 16 Jun 2008
Has thanked: 40 times
Been thanked: 38 times

Postby geizer » 04 Feb 2010

Hi flipflopper,

These are interesting points. Still don't auto trade, and far from implementing it but here are my 2 cents:
What I consider to be the true holy grail is the ability to efficiently use capital at all times while never risking more then a certain percentage at any given time. Without the ability to create strategies on the portfolio level it is impossible to know what is going on in the big picture at any given time.
Well said. The capital (Risk) management should be the cornerstone of any strategy.
Without it no strategy can guarantee me that I will stay liquid.
b) The strategy must always know what your current intraday and overnight buying power is so it can adjust positions to meet overnight requirements but is also able to take advantage of additional intraday margin.
This assumes the simultaneous intraday and swing trading. Not sure If I can handle it ... unless fully automated.
Involves different thinking, and appears to have different goals. Quite complex from the risk management point.
I choose to better focus on one, OR the other.
2. Can more accurately calculate slippage in back testing based on the bid ask spread, the avg daily volume, and the size of your position.
Not sure if understand what this for, but how economical would be an attempt to accurately measure slippage?
Both slippage and spreads also subject to how risk perceived by traders, and are functions of human emotions.
This alone creates enough uncertainty here. Unless you goal is to group symbols by estimating the slippage relatively, and do a comparative risk analysis, maybe?
3. When backtesting and optimizing a portfolio of instruments the results should allow you to drill down and see how each individual instrument performed. Also to group instruments that performed similarly together so you can easily identify which group of instruments works well with the strategy.
It's important, i think.
4. The ability to optimize based on Sharpe Ratio, Lowest Drawdown, least consecutive losers, etc. NetProfit is not the most important factor.
Your post suggests that you know which factors more important. First coming to mind is the Risk/Reward ratio. If this is one of them, then we are back to point 1a) of your post which again, i think is important.

--
Pavel

aljafp
Posts: 184
Joined: 28 Oct 2005
Been thanked: 1 time

Postby aljafp » 04 Feb 2010

I would like the ability to separate Signal Generation from Trade Management.
Currently trading software designs merge these 2 functions into one giant strategy.

But it would be neat if we can have a strategy to just generate trade signals and hand off the signals to another entity to manage the trade. ( handle trade entry, sizing, stop loss and breakevens etc )

brodnicki steven
Posts: 407
Joined: 01 Jan 2008
Been thanked: 3 times

Postby brodnicki steven » 05 Feb 2010

I would like the ability to separate Signal Generation from Trade Management.
Currently trading software designs merge these 2 functions into one giant strategy.

But it would be neat if we can have a strategy to just generate trade signals and hand off the signals to another entity to manage the trade. ( handle trade entry, sizing, stop loss and breakevens etc )
** I used to do something similar in TS, using their "include systems" function, not sure if that's what you're referring to but I hope MC adds that capability soon. That way you keep signals separate from management and can mix-match different signal MM module combinations.

janus
Posts: 835
Joined: 25 May 2009
Has thanked: 63 times
Been thanked: 104 times

Postby janus » 05 Feb 2010

I would like the ability to separate Signal Generation from Trade Management.
Currently trading software designs merge these 2 functions into one giant strategy.

But it would be neat if we can have a strategy to just generate trade signals and hand off the signals to another entity to manage the trade. ( handle trade entry, sizing, stop loss and breakevens etc )
That's almost what I'm doing already using DLLs. I've written my complex trading rules in another programming language and use MC signal studies to manage and execute the orders (entries, stop losses, profit taking, scaling in and out, etc.). However, I'm only interested in doing this in real-time - I've rarely performed back testing with MC. I back test using my own home grown software to be more flexible, which is what I used to do before I bought MC. In fact my application does play back too such that it doesn't display the future results as you scroll back and forward the charts. Has other useful features. Took me a long time to develop it but it was fun at the time. I almost added code to place the orders to IB using their APIs but decided to take the easier option and use MC to do this work. I just got too tired of programming and wanted to focus on trading. It would be nice to do everything in the one package but rarely does one find a single application does everything you need. Still I might one day as MC evolves. For now I'm happy with the mixed approach. In fact the more I think about it the more I like to have all my rules external to MC and just use MC as a trade execution engine. One reason for this is I have mixed feelings with EasyLanguage. But then that's just a personal view, which is colored by many years of programming experience in many other languages. Each to their own.

aljafp
Posts: 184
Joined: 28 Oct 2005
Been thanked: 1 time

Postby aljafp » 07 Feb 2010

I would like the ability to separate Signal Generation from Trade Management.
Currently trading software designs merge these 2 functions into one giant strategy.

But it would be neat if we can have a strategy to just generate trade signals and hand off the signals to another entity to manage the trade. ( handle trade entry, sizing, stop loss and breakevens etc )
That's almost what I'm doing already using DLLs. I've written my complex trading rules in another programming language and use MC signal studies to manage and execute the orders (entries, stop losses, profit taking, scaling in and out, etc.). However, I'm only interested in doing this in real-time - I've rarely performed back testing with MC. I back test using my own home grown software to be more flexible, which is what I used to do before I bought MC. In fact my application does play back too such that it doesn't display the future results as you scroll back and forward the charts. Has other useful features. Took me a long time to develop it but it was fun at the time. I almost added code to place the orders to IB using their APIs but decided to take the easier option and use MC to do this work. I just got too tired of programming and wanted to focus on trading. It would be nice to do everything in the one package but rarely does one find a single application does everything you need. Still I might one day as MC evolves. For now I'm happy with the mixed approach. In fact the more I think about it the more I like to have all my rules external to MC and just use MC as a trade execution engine. One reason for this is I have mixed feelings with EasyLanguage. But then that's just a personal view, which is colored by many years of programming experience in many other languages. Each to their own.
Similarly I don't do backtesting.

I feel that powerlanguage is great for programming indicators and generating trade signals fast.

I need more flexibilities like setting up bracket orders ( targets and stop losses, GAT orders) , setting breakevens, real time position sizing, using signals from one contract to trade another market etc.

I think the combination of MC as signal generation and NT as trade management would be great.

So my setups are the exact opposite to yours.
I use MC to generate Trade signals and use my own custom develop C# program to manage trades/positions with IB.

MC is a great charting/backtesting software with some auto trade functionality.
NT on the other hand is a great trade entry tool with improvement in strategy testing.
And NT is getting better and better with strategy development.
So MC needs to improve on the auto trading front or it will lose out.


Return to “MultiCharts”